Trump Seeks Curbs on GPT-5.6 as US Tightens Control Over Advanced AI Models

The incoming powerful AI model from OpenAI has been requested by the Trump administration to be released in stages. The move was made approximately two weeks after competitor Anthropic PBC, in response to regulatory pressure, pulled its most competent products from sale.

The US government asked OpenAI to distribute GPT-5.6 to a small group of trusted partners before making it available to the public, according to OpenAI CEO Sam Altman's statement to workers on June 24. Altman expressed growing US government anxiety regarding the capabilities of the most advanced models during the meeting with personnel. Altman emphasised that employees must collaborate with the Trump administration to address any concerns or recommendations on the security and limitations of OpenAI's forthcoming models.

US Government Putting Scanner on AI Firms

According to a White House official, the current administration is still working with cutting-edge AI research facilities to find common solutions to the problems caused by expanding the technology. Mythos 5 and Fable 5, Anthropic's two most powerful artificial intelligence programmes, were blocked from access by the world earlier this month.

The move was made in response to a government request that the corporation implement measures to prevent individuals from outside the US and foreign nationals from using the models, under the guise of national security concerns.

In an attempt to regain access to the models, Anthropic has subsequently engaged in discussions with the US government. For some in the field, the historic US government action was a wake-up call because it demonstrated the Trump administration's apparent eagerness to impose control over the rapidly growing AI industry.

Similar to Anthropic, OpenAI has been in a mad dash to release new models with superior performance in a variety of domains, including cybersecurity and computer programming. Workers at OpenAI are worried that the government's moves against Anthropic may hinder the ChatGPT maker's capacity to distribute its technology extensively.

According to Anthropic, the US government issued the order after they found a way to "jailbreak", or bypass, the guardrails of Fable 5, a newly released version of Mythos that they had previously prevented from doing cybersecurity chores.

Confusion Prevailing Among AI Players in USA

At the moment, AI labs are on their own to figure out ad hoc ways to notify the government about forthcoming sophisticated AI models. It is unclear who in government should be informed before launching a new AI model, when to inform them, how much information to disclose, and how far in advance to launch.

President Trump signed an executive order earlier this month, which some players in the AI sector are hoping will clarify the process. A voluntary framework has to be developed by the Trump administration and AI companies within 60 days of the directive's signing. Among other things, this framework grants the government access to so-called "frontier" models for a period of up to 30 days prior to their scheduled release.